Insectes soc. 45 (1998) 445–456 0020-1812/98/040445-11 $ 1.50+0.20/0 © Birkhäuser Verlag, Basel, 1998 Insectes Sociaux Research article Patterns of buzz running, a pre-swarming behavior, in the Neotropical wasp Parachartergus colobopterus V.O. Ezenwa 1, M.T. Henshaw, D.C. Queller and J.E. Strassmann* Department of Ecology and Evolutionary Biology, Rice University, P. O. Box 1892, Houston, TX 77251, USA, e-mail: [email protected] 1 Current address: Department of Ecology and Evolutionary Biology, Princeton University, Princeton, NJ 08544-1003, USA Key words: Parachartergus colobopterus, buzzing run, parasite dance, swarming, microsatellites. Summary Knowing whether workers or queens control reproduction is vital for understanding social evolu- tion in insects. Among epiponine wasps, a significant part of reproduction is the production of daughter swarms. One of the first behaviors indicative of swarming is the buzzing run – a behav- ior in which a wasp runs rapidly back and forth on the natal nest, bumping into nestmates while traversing a significant portion of the nest. In this study we investigated whether the buzz runners in Parachartergus colobopterus were queens who were not reproducing on the current nest (repro- ductive losers), or workers, as evidence from other species indicates. We suspected that loser queens, who may be anxious to swarm in order to increase reproductive output, could be the buzz runners because P. colobopterus swarms are primarily composed of queens who are not the mothers of the accompanying workers. Our results show that workers are the buzz runners, and that queen number does not influence the level of buzz running on a colony. We also found that buzz running is more common on large colonies with low worker relatedness, the sort that are like- ly to produce swarms. However, buzzing runs also occurred sporadically on smaller nests. The small nests had more nest parasites (mostly flies) than the large nests, and the parasite dance, which is very similar to the buzzing run, was seen only on these nests. Introduction Social insect colonies are more likely to fit the superorganism paradigm if collective worker interests control key features of colony reproduction (Queller and Strass- mann, 1998). Here we focus on one key stage of reproduction: new colony produc- tion. Successful reproduction in social wasps requires the production of new colo- nies in addition to the production of males and queens. In the subfamily Polistinae, * Author for correspondence. 446 Ezenwa et al. there are two main strategies by which new colonies are founded – independent founding and swarm founding (Jeanne, 1980). In independent-founding species, one or more inseminated queens initiate the colony without the aid of workers. In swarm-founding wasps the colony is started by a swarm composed of both workers and queens, and workers outnumber queens (Richards and Richards, 1951; Forsyth, 1978; Richards, 1978; Jeanne, 1980; West-Eberhard, 1982). Swarming is a coordinat- ed and synchronous migration of individuals in which the swarm follows chemical trails laid down by scouts to the new site (Naumann, 1975; Jeanne, 1981). The swarm-founding wasps include all of the nearly 200 species of the tribe Epiponini, the dominant eusocial wasps in the neotropics (Jeanne, 1991a and references therein). Large, multiple-queen colonies produce swarms (Richards and Richards, 1951; Solís et al., 1998; Strassmann et al., in press). The colony cycle begins when many workers and queens form a swarm and relocate to a new nest site. The queens in the swarm are rarely the mothers of the accompanying workers (Strassmann et al., 1998). This is either because they did not reproduce on the original colony (i.e, they were reproductive losers) or because the original colonies had such large numbers of queens that workers were unlikely to be daughters of the few queens joining the swarm. On the new nest the queens begin to produce workers and the nest grows. New queen production does not usually occur, however, until queen number is sub- sequently reduced to one (West-Eberhard, 1978; Strassmann et al., 1991; Queller et al., 1993). New swarm production is independent of queen production and occurs when colonies have many workers (Strassmann et al., 1998). This separation allows for new colony production to occur when it is most ecologically advantageous (Strassmann et al., 1997; Strassmann et al., 1998). The independence of new colony production and queen production allows for a certain flexibility in colony initiation, making it difficult to predict which group on the nest sets the timing of swarming. The workers in the colony, who possess in- formation on prey abundance, nest site availability, and worker-to-larvae ratios can use environmental and social cues to determine the optimal time for swarming. Also, if there are queens who are reproductive losers in the original colony they may have an interest in provoking swarming in order to increase their chances of repro- ducing. However, worker and queen interests in swarming will not necessarily be in conflict since neither party is likely to seek to initiate a swarm that contributes to either the new or the original colony’s failure. Prior to reproductive swarming there are definitive changes in wasp behavior which can be used as indicators that a colony is about to produce a swarm. The most widely observed of these behaviors is the buzzing run (Naumann, 1970; Jeanne, 1975; Forsyth, 1978; West-Eberhard, 1982). A buzzing run is characterized by fran- tic, jerky running by one to many wasps on the nest. It resembles the parasite alarm behavior (Naumann, 1970; West-Eberhard, 1982) which can trigger nest evacuation. The parasite alarm is a jerky movement performed in response to parasitoids such as flies, moths or wasps that lay eggs in the nest, producing young that devour the nest brood (West-Eberhard, 1969; Strassmann, 1981). This behavior occurs in many social wasps including non-swarming genera such as Polistes (West-Eberhard, 1969; Strassmann, 1981). Buzz running is also similar in form and function to the “Schwirrlauf” of Apis honeybees (Naumann, 1970; West-Eberhard, 1982), which is Buzz running in Parachartergus colobopterus 447 performed by scouts and incites the colony to swarm (Wilson, 1971). These simila- rities and other indicators suggest that the buzz runners on a wasp colony are asso- ciated with the induction of swarming. West-Eberhard (1982) described buzz running as the most widespread and characteristic pre-swarming behavior described by observers. The behavior has been observed occurring from at least a week before swarming up to the actual day of the event (Naumann, 1970; Forsyth, 1981). Naumann (1970) assigned buzzing run activity to the worker caste in Protopolybia acutiscutis. Forsyth (1978) marked wasps from Polybia occidentalis nests and found that active foragers were the in- dividuals exhibiting the buzzing run (breaking run) behavior. Observations by West-Eberhard (1982) also demonstrated that in Metapolybia aztecoides and Synoeca surinama the majority of wasps seen buzz running were old foragers. In M. aztecoides some of the buzz runners were younger workers, but none were queens. Although the evidence indicates that workers, specifically foragers, are the individuals performing buzz running, queens may also display the behavior in an attempt to facilitate swarming, particularly if they might reproduce more on a new colony then they did on the original one. In this study, we determined whether workers or queens are the buzz runners in Parachartergus colobopterus in order to shed light on who initiates pre-swarming behaviors in this species. Also, in order to learn more about when buzz running occurs on a colony and to further investigate the relationship between buzz running and swarming, we compared the frequency with which the behavior occurred on large, potentially swarming colonies to that on small colonies where swarming was unlikely. Finally, to examine the effects of buzz running on colonies, we studied the responses of other colony members to buzz runners. We conducted all of our studies in the rainy season when swarming is most frequent in this species (Strass- mann et al., 1997 and unpublished observations). Materials and methods Videotaping and collection We studied 10 Parachartergus colobopterus colonies in Maracay, Venezuela (10° 16¢ N 67° 36¢ W, altitude 445 m) at the Universidad Central de Venezuela in the middle of the wet seasons of 1993 and 1995. We studied the 5 small colonies (< 250 wasps: V20-8, V20-12, V20-14, V20-38, V20-42) from 29 July to 6 August 1993, and studied the 5 large colonies (>340 wasps: V21-1, V21-2, V21-8, V21-13, V21-18) from 29 July to 6 August 1995. First, we marked as many individuals as possible on all 10 nests with dots of Testors© enamel over two mornings. Then, to videotape each nest we pulled back the nest envelope in order to expose the combs. We videotaped each colony over two days so our samples of behavior provide a detailed picture of a brief period of time. Upon completion of videotaping, we collected all of the colonies (the nests and the wasps) and stored the wasps in liquid nitrogen or dry ice for transport back to the lab for photographing, dissection, and genetic analysis. 448 Ezenwa et al. Behavioral observations In order to facilitate the identification of specific individuals on tape, we photo- graphed and numbered each wasp collected. We watched videotapes from all nests for instances of the buzzing run behavior, as well as instances of the parasite alarm/dance because of its similarity to the buzzing run. We identified as many of the actors as possible by their marks. We watched a total of 137 hours of tape; hours of tape per nest ranged from 10 to 18 hours, with an average of 13.7 hours per nest.